The inclusion of human factors (HF) in mathematical models is proving crucial to allow complex driving behaviour and interactions to be explicitly considered to capture driving phenomena. An important area where such integration is required is for the role of anticipation by drivers to compensate for critical traffic situations. In this paper, we introduce the concept of Anticipation Reliance (AR), which acts as a demand lowering compensative effect for the driving task by relying more on anticipation. We implement AR in a generic multi-scale microscopic traffic modelling and simulation framework to explore and explain the effects of HF on traffic operations and safety in critical traffic situations. This concept addresses a disparity in the description of driver workload in relation to the execution of driving tasks in regard to the confidence that drivers place on tasks that are sub-consciously catered for. The crossover from HF to a mathematical description of this role of AR introduces a ground-breaking concept that explains and models the mechanisms that allow drivers to compensate and avoid accidents in many circumstances, even when driving errors or sub-optimal driving performance occurs. By and large, the HF effects can be subdivided in effects on perception and anticipation; effects on sensitivity and response to stimuli; and effects on personal attributes and characteristics. A key aspect of the framework are two intertwined driver-specific mental state variables—total workload and awareness—that bridge between classic collision-free idealized models for lane changing and car following, and HF models that explain under which conditions the performance of drivers deteriorates in terms of reaction time, sensitivity to stimuli or other parameters. In this paper, we focus on the awareness construct, as described by AR, and explore the effects. We prove the effectiveness of the approach with a case example that demonstrates the ability of the model to dissect a complex traffic situation with both longitudinal and lateral driving tasks, while endogenously considering human factors and that produces accident avoidance and occurrence within the same order of magnitude compared to real traffic accident statistics.

实践证明，在数学模型中包含人为因素（HF）对于允许明确考虑复杂的驾驶行为和相互作用以捕获驾驶现象至关重要。需要这种集成的一个重要领域是驾驶员的预期作用，以补偿关键的交通状况。在本文中，我们介绍了预期依赖（AR）的概念，它通过更多地依赖预期来降低驾驶任务的需求补偿作用。我们在通用的多尺度微观交通建模和仿真框架中实施AR，以探索和解释HF对紧急交通情况下交通运营和安全的影响。该概念解决了关于驾驶员对下意识地迎合的任务的信心的驾驶工作量的描述与驾驶任务的执行之间的差异。从HF到对AR的这种作用的数学描述的引入引入了突破性的概念，该概念说明和建模了使驾驶员能够在许多情况下补偿和避免事故的机制，即使在发生驾驶错误或次优驾驶性能时也是如此。总的来说，HF的影响可以细分为对感知和预期的影响。对敏感性和刺激反应的影响；以及对个人属性和特征的影响。该框架的一个关键方面是两个相互交织的驾驶员特定的心理状态变量（总工作量和意识），它们在经典的无碰撞理想化模型（用于换道和跟随汽车）之间架起了桥梁，而HF模型则解释了驾驶员在何种情况下性能会下降在反应时间，对刺激的敏感性或其他参数方面。在本文中，我们将重点放在AR所描述的意识构造上，并探讨其影响。我们通过一个案例来证明该方法的有效性，该案例演示了该模型分解纵向和横向驾驶任务的复杂交通情况的能力，同时内生地考虑了人为因素，并在相同的数量级内避免和发生事故与实际的交通事故统计数据相比。